Has social media really led to diversity, or are we all just sharing the same few links?

While we can argue without doubt that social media has lead to a proliferation of content, through ease of publishing, cheaper computers and improved internet capability, when you look at whether social media has lead to diversity of consumption of content, the point becomes harder to make. Sure, there is a lot more information ‘out there’ , but when you consider the fact that the majority of internet traffic is concentrated on (relatively) few sites, we need to investigate further to see if the internet and social media is as democratic as we initially thought.

Social platforms in particular are often accused of being echo chambers. That is to say that instead of a wide range of links being shared, what we have instead is a situation in which we have built close communities of shared contacts, where the same information and links are being perpetually shared and discussed. But is this really the case, or is it too primitive a view of how social media functions?

The challenge from Facebook

Eytan Bakshy, who works on the data team at Facebook, has just released the findings of his team’s lengthy study into the accusation that Facebook is really just an echo chamber. He found that while it is certainly the case that we are more likely to find links that come from our ‘strong’ ties – i.e. those few, closest contacts of ours on the platform, our interest in new links shared by ‘weaker’ ties does not diminish.

The likelihood of finding content and the eventual interest in it are two very different things. This is also explored further in the likelihood to share that content on your Facebook wall. Bakshy found that on average, people were ten times more likely to share links they found through a ‘weak tie’, whereas they were just six times more likely to share content posted by ‘strong ties’.

This, for me, presents a bit of a problem. It seems we are likely to have a higher interest in – and propensity to share – content that is from our weaker ties, but by their very nature, we come into contact less with these weaker ties. Essentially what our social interactions and communities rely on, is that which we are exposed to least. We all want to be the one to share something new with our communities, but coming into contact with that link is difficult. Paradoxically, our weak ties become more influential, refuting the idea of social media being an echo chamber.

Merging communities

Where we can perhaps see a threat to social media encouraging diversity, is in the very fact that it is growing exponentially and becoming increasingly prevalent in society. This has resulted in fewer disparate communities than we might once have had in the earlier days. When different platforms first emerged, we were relatively adept at maintaining different communities on different sites, with different community members and subsequently different information being shared. But as cross-platform integration has emerged, this is decreasing.

Where Twitter updates once stayed on Twitter, they now feature on Facebook. Where a blog comment stayed on a blog, it can now be published with your Facebook profile. In this way the content between different platforms is becoming less distinct : an unfortunate and unexpected outcome of the growth of social media.

The sources for a particular story or link may have increased as a result of merging communities, but are the actual links any different? And if they’re not, what does this then mean for the concept of the internet being a democratising force?

Our natural filters

The problem is of course, that the internet can’t negate our natural desire to navigate to the content that we find more familiar and comfortable, whether this is through going to the content directly, or it seeming more familiar and comfortable because we’re navigating to it through a link shared by a close contact.

This is discussed by Cass R. Sunstein in Republic.com 2.0 where he looks at how we narrowly filter the information we receive, choosing only (or largely) that which we want to hear. While this may underestimate human nature and curiosity slightly, it is understandable that we find comfort in the familiar and may find this more accessible by default i.e. through the sophistication of language used, references made, alignment with our political or religious beliefs etc.

When we don’t have much time online, it’s understandable that we filter through content to prioritise those sources that we already trust. Again it leads back to the problem that while social media has undoubtedly lead to a proliferation of content, this content isn’t necessarily being very well discovered. We’re all busy publishing, but who’s actually reading?

The threat of social search

In this way, it is possible to see the advent of social search as a threat to the democratization of information. While this is currently very much in limbo as we look to Google to perfect social search, who are failing pretty miserably, it could have profound effects on, and indeed contribute to, the echo chamber. Right now when we put a search query into Google, we are returned with the results that are deemed most relevant to that query.

And while some efforts have been made to personalize results, largely we are all pretty much seeing the same results (location depending etc..). If social search is to develop however, the theory obviously goes that we will see the results most relevant to US, and this may be very different to the information that is most relevant overall. It means that the narrower field of link discovery will exacerbate, as the search engines do the jobs of link discovery that is currently being done by our social communities.

If social search is to avoid the echo chamber effect, there must be a careful balance between the content within our communities, and that content that we might never normally come across, yet which relates to what we want to find.

And just as the route to a democratic approach for discovering content lies with the search engines, it also lies with the social platforms themselves. This brings us back to Facebook’s study by Bakshy, as he looked at the role of EdgeRank in content discovery. Whereas EdgeRank typically functions by prioritizing the content from our stronger connections or ties, if it instead promotes that content that we were less likely to discover from closer ties, it allows for more diversity and increased sources for information.

There is still a lot of learning to be done by social networks here, in order to get the balance right. We will find comfort (and assume usefulness) in a social network if we see the familiar – i.e. our closer connections, but the end result (and actual usefulness) may come from discovering content from lesser known connections, that we wouldn’t naturally filter ourselves. In this way, to continue democratization of information and less concentration of traffic on a few privileged sites, we need the social networks to do the sophisticated filtering for us. This is the role that social platforms now play. Where they started out as facilitating production of content, they now need to consider how that content is really distributed.

Some may question whether this is important, or why we need to constantly find new sources of information. It is important however, if we are to become less separatist and to begin to understand things within a wider world view. Knowledge of the unfamiliar can only be a good thing, and can contribute to a more connected, knowing society that is ultimately less disparate. It is also important to avoid the concentration of power that we’ve seen with traditional media. We all have the ability to publish to the world but that only means anything if people are actually reading what we write.